Measurements and Modeling of Frost Depth in Railway Tracks

In Finland, increases in the maximum allowable operating speed on railways have led to stricter smoothness requirements for rails. Despite continuous rehabilitation of the track structures, decreased speed limits have been required due to seasonal frost action. To improve the understanding of frost action phenomena in railway structures, a real-time in-situ monitoring system was installed at 14 sites within the Finnish railway track network for the purpose of measuring temperatures in the track structures and vertical displacements of selected railway sleepers. Based on 5 years of data collected from 2009 to 2013, during which time the maximum frost depth was observed to range from 0.9 m to 2.4 m, multivariable regression models for predicting the maximum frost depth at the instrumented sites were developed. Several factors were shown to influence the maximum frost depth, including latitude, longitude, air-freezing index, rainfall, elevation, and ditch depth. With comparatively high R² values of 0.73 and 0.87, the resulting regression models may provide satisfactory utility for predicting maximum frost depth at sites similar to those that were instrumented for this research. When considered with other information such as track structure thickness, predictions of maximum frost depth can be useful for development of site-specific reasons for observed frost problems in railways.